Preservation of scores of the quality of traffic to network sites across clients and over time

Abstract

A software and/or hardware facility for scoring the quality of traffic to a site accessible via the Internet or other network. The facility may generate training set data and use the training set data to identify parameters indicative of fraudulent traffic to a site and reduce the effect of fraudulent traffic advertisers and publishers. The facility may score the quality of traffic and determine combinations of parameters that are indicative of the quality of traffic to the site. Traffic to the site may be scored based on the combination of parameters associated with the one or more sessions. Lower scores are indicative of traffic having little value to a publisher, advertiser, or third party; higher scores are indicative of traffic having greater value.

1 . In a method of generating scores for a quality of traffic received by a network site, the scores representative of likelihood that the traffic is of value to the network site, a method of preserving the scores of the quality of traffic represented by at least two data object sets having at least one overlapping rule set, the method of preserving scores comprising:
 receiving a first data object A having a first rule set and a second data object B having a second rule set;   determining an intersection of the first and second rule sets AB;   building an association table of result vectors F A/AB  for the first data object A and first rule set and the overlapping rule set;   determining odds of conversion and average scaled scores for each result vector f of the association table F A/AB ;   estimating average scores based on odds of conversion and spread of association table F A/AB ;   building an association table of result vectors F A/AB  for the second data object B and the second rule set and the overlapping rule set AB;   retrieving from table F A/AB  odds of conversion and average scaled score for each result vector f of association table F B/AB ;   estimating average scores based on odds of conversion and spread of association table F b/AB ;   estimating average scores based on odds of conversion and spread of the second data object B;   scaling raw scores of the second data object B using average scores based on odds of conversion and spread as core parameters.   
     
     
         2 . In a system for generating scores for a quality of traffic received by a network site, the scores representative of likelihood that the traffic is of value to the network site, a system for preserving the scores of the quality of traffic represented by at least two data object sets having at least one overlapping rule set, the method of preserving scores comprising:
 a receiving component for receiving a first data object A having a first rule set and a second data object B having a second rule set;   a rules intersection determining component for determining an intersection of the first and second rule sets AB;   a first table building component for building an association table of result vectors F A/AB  for the first data object A and first rule set and the overlapping rule set;   an odds of conversion and average scaled scores determining component for determining odds of conversion and average scaled scores for each result vector f of the association table F A/AB ,   a first estimator for estimating average scores based on odds of conversion and spread of association table F A/AB ;   a second table building component for building an association table of result vectors F B/AB  for the second data object B and the second rule set and the overlapping rule set AB;   a retrieving component for retrieving from table F A/AB  odds of conversion and average scaled score for each result vector f of association table F B/AB ;   a second estimator for estimating average scores based on odds of conversion and spread of association table F B/AB ;   a third estimator for estimating average scores based on odds of conversion and spread of the second data object B;   a scaling component for scaling raw scores of the second data object B using average scores based on odds of conversion and spread as core parameters.   
     
     
         3 . A method of generating scores for a quality of traffic received by a network site, the scores representative of likelihood that the traffic is of value to the network site, a method of preserving the scores of the quality of traffic represented by at least two scoring time periods having an overlapping scoring time, the method of preserving scores comprising:
 Determining, an average score based on odds of conversion and spread for a first scoring period A;   estimating, an average score based on odds of conversion and spread for the overlapping period AB using available scaled scores from period AB;   modifying the multiple rule set applied by hardware computer processing to the first scoring period A to generate a modified rule set to be applied by hardware computer processing to a second scoring period B;   applying the modified rule set in the second scoring period B to generate raw scores for the second scoring period B; and   scaling, raw scores for the second scoring period to maintain the average score based on odds of conversion and spread of the overlapping scoring period AB.   
     
     
         4 . In a system for generating scores for a quality of traffic received by a network site, the scores representative of likelihood that the traffic is of value to the network site, a system for preserving the scores of the quality of traffic represented by at least two scoring time periods having an overlapping scoring time period, the system of preserving scores comprising:
 a score determining component for determining an average score based on odds of conversion and spread for a first scoring period A;   an estimator for estimating, an average score based on odds of conversion and spread for the overlapping period AB using available scaled scores from period AB;   a rule modifier for modifying the rule set applied to the first scoring period A to generate a modified rule set to be applied to a second scoring period B;   a raw score generator for applying the modified rule set in the second scoring period B to generate raw scores for the second scoring period B; and   a raw score scaling component for scaling raw scores for the second scoring period to maintain an average score based on odds of conversion and spread of overlapping period AB.